Advice regarding filters

Wed, 11/01/2017 - 21:57

I have recently become interested in Photometry. I currently have an imaging setup with LRGB filters. I plan to change to photometric filters if that becomes my main focus. 

My question is whether, in the meantime, it is acceptable to submit RGB data as transformed TG TR and TB as if from a DSLR. I have read carefully both the DSL and CCD AAVSO manuals and understand the limitations of the TG TR AND TB system. I cannot find any data on the bandwidth of DSLR  RGB to see how comparable my astronomik RGB are, but certainly my "TG" results compare well with AAVSO results in both TG  and V. 

I have calculated transformation coefficients for my filters (see below). I dont know how these compare with DSLR coefficients and whether that gives any further indication as to suitability.

35 standard stars NGC 7790

Tbv  1.314 +-0.013

Tvr  1.242+-0.008

T v_bv  -0.0231 +-0.006

T b_bv 0.216 +-0.01

T r_vr -0.242+-0.001


My second question is what would be the current recommendation for photometric filters as and when i do change. I have read various threads about Sloan filters, but on the last discussion i found UBVRI still seemed the preferred option. Is this still the case?

thanks for any assistance you can give


Any response to this? I

Any response to this? I assumed there would be an AASVO view. 

Is it reasonable to submit CCD RGB data treated in the same way as DSLR data? I am aware that stars with significant emission lines must be avoided. 

I hope this is not a stupid question!


American Association of Variable Star Observers (AAVSO)
DSLR like photometry

Hi Mattew,

I think this is certainly possible but what type of LRGB filters are you using ? The RGB filters of DSLR are not of "rectangular" response contrarily to many LRGB filters used for astro imaging (some, for true colors imaging, are more similar to the DSLR ones). If your filters have such "rectangular" response your case could be significantly different than DSLR. DSLR filters are designed for providing a full color space comparable to such of the human vision. By the way the response of those filters have specific shapes with long roll-off slopes. The DSLR also include a strong IR rejection using both tinted glass and interferential layers, be carefull about that, the IR leaks could afffect the photometry of stars having strong IR. The typical "rectangular" LRGB filters are not able of true colors imaging as the wavelengths inside a given channel can't be differentiated. 

You could find an analysis of the color response of DSLR filters in my VSF paper in JAAVSO:

I could make a similar simulation for the RGB of LRGB filters if data are available.

DSLR RGB channels are able of very good V photometry after transformation an better with the VSF technique, B and R are more questionnable depending star spectral types, this is also due to the low SNR of the R channel which is not your case. 

Attached is a comparison of various band-pass definitions, you could see that the G response of the DSLR is very similar to the VT one of Tycho (Tycho 2 catalog, very  usefull for us). 

Clear Skies !


American Association of Variable Star Observers (AAVSO)
Astronomik LRGB

Matthew, I  found two types of RGB filters at Astronomik, I think the one you use is the LRGB typ2C. This one is not having the classical "rectangular" response and its green  response shape is very near the Johnson's V and  better centered than the DSLR G one. This explains your Tv-bv is so low: -0.0231. The coef. for a DSLR G is more like 0.12 ~0.14. Your Vj using that filter shall be very good with very little transformation ! I even suspect it's as good as some of the commercial Johnson's filters... 

The B and R responses of the typ2C are not as good as the G one but anyhow much better than the DSLR ones. 

I think this filter set is very interesting for photometry and you have no urgence to buy those expensive Johnson/Cousins filters !

Clear Skies !



RGB submission

Hi Matthew,

You are taking the right approach - transforming your RGB into Rc/V/B.  Your coefficients seem reasonable to me.  You can certainly submit the transformed magnitudes to the AAVSO.

The transform ONLY works within the color range of the cluster you used.  Where this kind of transformation fails is for very blue (pulsating B stars) or very red stars (like Miras).  Your best bet is to start with variables with "normal" colors, like cepheid/RR Lyr and eclipsing finaries.

If you get photometric filters, I still recommend the Johnson/Cousins filters for now.  Your first two filters should be B&V, with Ic, Rc, U as your next choices in that order.  There are more and more campaigns from professionals who would like Sloan magnitudes, but I think you should "get your feet wet" with the J/C system first and then talk to the researchers directly if you want to participate in one of the Sloan campaigns.  If you demonstrate your abilities, sometimes professionals will even buy the necessary filter for you.

Good luck, and welcome to the photometric club!


Roger and Arne

Roger and Arne

thanks to both of you for your response. 

Roger - my RGB filters are Astronomik and do indeed have the rectangular or flat profile that you refer to, as shown in their published specification though more so in B and R (see attached file). I now realise from your reply that this could be significant. Thanks for the paper, which I need to read and digest (and even then I suspect a lot may be over my head). I would not think that it would be worth the effort of setting up your  method for LRGB filters, as I guess their use is going to be much more limited for obvious reasons.

Arne - thanks for the encouragement. Working through the processes with non photometric filters, including calculating transformation coefficients by spreadsheet is a good way of starting to understand, even if practical use is limited. The decision to switch to photometric filters will be quite a step as I run a remote setup so I cannot switch back and forth too easily. But I can see it happening as the bug continues to bite!






Astronomik RGB


I missed your last post where you have also uploaded the Astrronomik profile. Certainly my results with the G filter are v close.

Thanks for your help